Thomas Coombs is an accountancy and business advisory firm providing accounting, tax, audit, payroll, corporate finance, and advisory services. Its work creates a familiar challenge for modern accountancy firms: the more clients, services, data sources, and reporting cycles there are, the more manual finance operations can quietly absorb capacity.
We helped Thomas Coombs approach that problem as an automation and workflow design challenge, not a headcount problem. The goal was to increase finance delivery capacity, reduce repeated manual work, and give teams clearer operational visibility without adding cost at the same rate as volume.
The problem
Finance and accountancy workflows are full of repeatable but high-judgement work: collecting records, checking documents, reconciling transactions, preparing management information, spotting exceptions, routing queries, and keeping client teams updated.
At scale, those workflows create hidden drag. Skilled people spend time moving data between systems, chasing missing information, checking routine exceptions, formatting reports, and reconstructing workflow status from spreadsheets, inboxes, and disconnected platforms.
Thomas Coombs needed a way to make finance operations more scalable while keeping the controls, auditability, and professional judgement that clients expect from an advisory firm.
What we changed
AI-assisted document processing: We used AI to classify, extract, and structure finance documents so invoices, records, statements, and supporting files could move into review workflows with less manual keying and fewer repeated checks.
Automated reconciliation workflows: We created rule-based and AI-assisted matching workflows to compare transactions, supporting records, account movements, and known exceptions, reducing the amount of manual review needed for routine reconciliation work.
Exception detection and routing: We designed workflows that surface anomalies, missing information, duplicate records, out-of-pattern movements, and low-confidence matches so skilled reviewers can focus on the cases that actually need judgement.
Management reporting automation: We automated parts of the data preparation, validation, and report assembly process so recurring finance packs and dashboards could be produced with less spreadsheet manipulation and manual formatting.
Operational visibility layer: We created workflow views that show queue status, blockers, review stages, overdue items, and exception volumes so teams can manage finance delivery through live operational signals instead of manual status updates.
What this unlocked
Estimated 40% to 60% reduction in manual document handling: AI-assisted classification and extraction reduce the amount of time spent opening files, reading routine documents, and moving information into downstream workflows.
Estimated 50% to 70% faster reconciliation on covered workflows: Automated matching and exception routing reduce the amount of line-by-line checking required before a reviewer can focus on unresolved or higher-risk items.
Estimated 30% to 50% faster management reporting cycles: Automated data preparation, validation, and report assembly reduce spreadsheet-heavy production work around recurring finance packs.
Higher reviewer capacity without lowering control: Skilled finance staff can spend more time on judgement, advice, and client-facing work because routine classification, matching, routing, and formatting are handled by systems.
More scalable client delivery: Thomas Coombs can support more finance workflow volume without every increase creating the same increase in manual administration, query chasing, and reporting production effort.
Why this matters
Accountancy and advisory firms do not win by replacing professional judgement. They win by protecting it. The opportunity is to remove the repeated manual work around finance delivery so specialists can spend more time on review, interpretation, advice, and client outcomes.
Thomas Coombs shows how AI and automation can be applied inside serious finance workflows without treating them as generic back-office tasks. The systems need controls, exception handling, auditability, and clear human review points.
That is where the commercial value appears: more work can move through the firm, clients get faster and clearer service, and teams are less constrained by repetitive operational tasks.